combination of geographic information system (gis) and radio frequency (rf) planning software, to...

7/28/2019 Combination of Geographic Information System (GIS) and Radio Frequency (RF) Planning Software, to Assess the

1/8

Combination of Geographic Information System (GIS) and RadioFrequency (RF) Planning Software, to Assess the Market Potential ofWireless Broadband Internet to Unserved Rural Populations

Daniel Cossette*, M. Sawada*, Grald Chouinard**, G.Briggs***, Peter G. J ohnson****

*Laboratory for Applied Geomatics and GIS Science (LAGISS), Department of Geography, University of Ottawa,60 University St., Ottawa, ON K1N 6N9geomatics.uottawa.ca**Program Manager, Rural and Remote Broadband Access, Communications Research Centre (CRC) Canada.***Industry Canada****Chair, Canadian Polar Commission and Department of Geography, University of Ottawa

ABSTRACT

Millions of Canadians do not have broadband Internet access and the majority of these households reside withinrural and remote communities of Canada. This situation has led to an urban/rural broadband divide. Barriers existin the deployment of wireline broadband for rural and remote regions because of the significant expense to installand maintain a wired infrastructure required to reach remote dwellings and businesses. In such situations the costs

outweigh the benefits of wireline access. Spatial differences in Internet accessibility have led to an urban/ruralbroadband divide in Canada.

Accessibility to terrestrial broadband wireless technology is a question based on spatial and regional topography andthe capacity for different technological solutions to reach profitable population bases. Various standardsorganizations such as the IEEE (Wireless Metropolitan Networks WMAN 802.16, Wireless Regional AreaNetworks WRAN 802.22) are developing technologies that would permit wireless Internet access over much greaterdistances than current technologies. The benefits of such new technologies include rapid deployment, flexibility,scalability and the lack of a physical connections being required to connect a subscriber.

Our approach to determine potential market served by news broadband wireless technologies in Canada combinesRadio Propagation software (Planet EV) and Geographical Information Systems (GIS). We demonstrate a methodcapable of identify areas where deployments would be profitable from a business viewpoint and areas where

difficulty would arise either due to topography, vegetation or lack sufficient households . This method has yet to befully realized however from preliminary results we found that a considerable amount of the unserved populationcould be provided with access. This study is of internet to service providers, telecommunication manufacturers andboth policy and frequency regulators in Canada.

Key words: Broadband, Internet, Remote Communications, Broadband Divide, Geographic Divide, WirelessTelecommunications, GIS, Radio Propagation, Wireless Access, WiMAX

1. IntroductionThe idea of an Information Society was hinted at in the early 1960s (Wellar, 1985) and presently, manyhouseholds are taking part. Prior to taking part in the information society, households with broadband internet firsthad to have broadband services available (Figure 1). Having access to broadband (high-speed) internet enables

households to take part in distance learning, e-government, e-commerce, banking, communication, cultural andother information services. Without broadband communications, households cannot enjoy the benefits of being partof the Information Society which creates a divide or an inequality households with and households withoutbroadband access. Many households are unable to part-take in the information society due to the geographic dividebecause broadband services are simply not available in their particular geographic location. Hence, a geographicdivide exists, because these households are located in an area where broadband is non-existent.

The reason why the majority of rural and remote areas lack broadband is there so few households to financiallysupport the required infrastructure needed. Unlike rural areas, urban regions have more households per unit area

1


2/8

resulting in fewer essential infrastructures per household because of the short distance between homes. This resultsin a polarized society of two extreme groups: those with access to technology and those without access (Fourieand Dowell, 2002), and in the case of broadband internet access, a so-called geographic divide leads to have andhave-nots.

ACCESS TOINFORMATION(BROADBAND)

DATA ANDINFORMATION

SOCIETY

KNOWLEDGESOCIETY

Figure 1. Schematic illustrating the steps from access to knowledge with regards to broadband wirelesstelecommunications.

Organizations such as the Institute of Electrical and Electronics Engineers (IEEE) are working on the developmentof wireless technologies that can cover greater distances than are currently possible (WiMAX). Benefits of suchnew technologies include rapid deployment, flexibility, affordability, scalability and the major benefit of not havingto physically hook up to homes. Aside, simply connecting millions that would otherwise be left out of theknowledge society the benefit of wireless technology is that it has the ability to allow connections that are non-line-of-site (NLOS) and service radii of up to 50km1. However, to facilitate a financially sustainable broadband tower, a

sufficient number of households capable of to a communication tower are required. In order to provide broadband,wireless technologies must overcome possible interferences to radio-waves.

Our approach combines Radio Propagation software (Planet EV) and Geographical Information Systems (GIS) inorder to determine potential market served by new broadband wireless technologies in Canada. We demonstrate amethod able to identify areas where deployments would be profitable from a business viewpoint and areas wheredifficulties would arise due to variable physical conditions affecting signals quality and strength. This method hasyet to be fully realized however from preliminary results we found that a considerable amount of the unservedpopulation could be provided with access.

2. DataThe importance of detailed data is vital in making accurate predictions. The idea is reflected in the acronym GIGO(Garbage-in Garbage-out). It is commonly known that higher resolution data yields better results (Thieken, 1999)

because low resolution data could potentially cause misinterpretations when using RF software at a local scale. Thedata used in our project are listed inTable 1.

Table 1. Data required and land use along with the resolution/scale and source.General Descript ion Type of Data Resolution / Scale Source

Digital Elevation Model(DEM)

Raster 30m DMTI Spatial Inc

Clutter Data(Land Use and Type)

Raster 30m and 90 m DMTI Spatial Inc

ProvincialBoundaries

Polygon 1:50,000 DMTI Spatial Inc

Census Dwelling CountDissemination Area

Polygon 1:50,000 Statistics Canada

Rivers andLakes Polygon 1:50,000 DMTI Spatial Inc

Communities with accessand without access (served /

unserved)Table n/a

Department ofIndustry

Block Population Point Variable Statistics Canada

1 Alvarion. Introducing WiMAX: The next broadband wireless revolution. 2004. (Date Accessed: 23 July 2004)http://www.alvarion.com/RunTime/Materials/pdffiles/Wimax_wp.pdf

2
http://www.alvarion.com/RunTime/Materials/pdffiles/Wimax_wp.pdfhttp://www.alvarion.com/RunTime/Materials/pdffiles/Wimax_wp.pdf


3/8

Clutter data is utilized in the wireless telecommunication industry in order to determine optimal locations of towers.This data is specifically developed for Radio Frequency (RF) propagation in order to refine signal loss predictionmodels according to the characteristics of the underlying terrain and ground cover (DMTI Spatial, 2005). This datais important because surface features can affect wireless communication and this data is classified as the man-madeand natural features that may impair radio frequency propagation by reflection, diffraction, absorption, or scatteringof the transmission waves (DMTI Spatial, 2005). The Digital Elevation Model (DEM) is also used for determiningthe RF propagation as it used for signal loss.

DMTI Spatial Inc. of Markham, Ontario created the clutter data from the Government of Canadas NationalTopographic Data Base (NTDB). They classified the data based on land use into both 30 and 90 metre rasterdatasets. The classification use was as follows: freshwater, ocean, coniferous forest, deciduous forest, open land,airports, industrial, urban, dense urban and wetland. One could create this data however it would be very timeconsuming if they would want to examine a large area.

The provincial boundaries are used to delineate an accurate outline of Canada, where the rivers and lakes are used tohelp eliminate the location where households can not exist. The communities with access and without access dataare provided by the Department of Industry. The data helps indicate which communities have broadband access inthat area and is spatial represented by being combined with the block population.

3. MethodologyAs aforementioned, the general method for the examining the successfulness of the providing wireless broadband tounserved households is to combine a Geographic Information Systems (GIS) approach with Radio PropagationAnalysis (RPA). Combining the two methods allows for factors that are known to cause signal degradation to beconsidered at the regional level while conducting an examination at the national scale. An illustrated diagram of theprocedure is outline in Figure 2.

Figure 2. Steps for the examination of broadband to households without broadband access.

3.1 Identification of RegionsThe first step of the analysis is to identify various vegetated and topographic regions of Canada in order toconduct RPA in these regions. This step identifies the various regions such as flat open land (SouthernSaskatchewan) verses hilly vegetated (Northern Ontario) areas. This step is required in order to determine ifbroadband is capable of being delivered in different regions of varying clutter and elevation.

The data provided by DMTI Spatial is provided by map sheet and therefore requires lengthy manipulationin order to have the data in one map sheet prior to identifying general regions. Additionally, because of theprediction model being used in the RPA, data was reclassified in three categories: open land, vegetated andurban area.

3.2 Radio Propagation AnalysisAfter having outlined various geographic regions, a RPA analysis is conducted in each of these regionsbased on criteria provided by the WiMAX Forum. WiMAX was selected over parallel technologies becausethe IEEE has a number of publications available for conducting a proper radio propagation analysis. For

3


4/8

this analysis both DEM and clutter data are in the Radio Propagation software called Planet EV, developedMarconi. Some of the variables used in the software include an examination at a distance of 50 kilometres(km), tower heights of 30 metres, a receiver height of 5 metres, frequency of 2500 MHz, and customizedpropagation models for urban, suburban and quasi-urban. The propagation models are based on research byHata. (Hata 1980)

3.3 Population Distribut ionIn the third step the dwellings counts are the primary inputs into the analysis. We assume that eachdwelling is a possible subscription unit to a wireless broadband service. Given that the spatial data for thecensus boundaries are general maps, the accuracy of the maps was increased by removing (cutting-out)water bodies and other areas where people and households cannot physically reside (Figure 3). While inreality households will tend to be clustered in particular parts of a census unit, the necessary assumptionmade here is that the values measured within a census unit are distributed equally across the census unit. Assuch, we distributed the households within and across units into a more organized and systematic grid.

Figure 2. Cartographic model representing the method used for preparing the population data.

The irregular shapes of the dissemination areas (DA) dwelling data from Statistics Canada were convertedinto discrete non-overlapping 1 x 1 km grid cells (Figure 3). The 1 km resolution produced more that 20million grid polygons and pushed the limits of the processing power for the eight computers used inanalysis2. The procedure of preparing the population is depicted in Figure 3 as a flowchart that outlines thelogic of the analysis process. The method used in distributing households from irregular DAs into regulargrid cells is as follows:

1. Join of Dissemination Areas with data provided by the Department of Industry to determine DAscurrently without broadband access, here calledDA Population;

2. Removal of the DAs without dwelling counts;3. Removal of areas where broadband is available;4. Removal of lakes and river areas of each DA;5. Calculation of the area of each DA, here calledArea of the DA;6. Creation of a 1 x 1 km grid with a unique identifier (id number) for each cell spanning all of

Canada.7. The grid created in step six is unioned with the layer representing unserved DAs, step 5 and 6.

This results in the creation of numerous polygons where more than one DA intersects a cell. Eachnew polygon contains the cell identification number and DA dwelling count.

8. The areas of all new polygons are calculated, here calledArea of Split Polygon;9. The population of each new polygon is calculated as:

2Dell Precision 650 Workstations each with two Xeon 3.2 GHz processors and 4 GB of RAM.

4


5/8

Area of theSplit PolygonNewPopulation DA Population

Area of theDA=

;

10. Using the unique identifier, the New Population is the summed for each cell identifier anddissolved into 1 x 1 km cells for subsequent analysis.

11.Total dwelling counts from step ten are compared to the unserved dwelling counts in step 1 forCanada to ensure process validity. Therefore, the number of dwellings at the beginning needs to bethe same as comes out.

Figure 3. Example of regular grid overlaid on dissemination areas (DAs); light grey areas are the DAs,darker grey areas are water bodies or DAs without populations that were removed from analysis; dottedlines represent the 1 km cells (one is highlighted in black with dimensions); one example is given of how

one cell intersects seven different DAs and as such, the number of households assigned to the cell is equalto the sum of households taken from the proportional population assigned to each piece of the 7 DAsintersected by the cell.

3.4 Minimum Sustainable SizeWith the households distributed within the 1 km2 grid cells, our next step was to estimate the minimumnumber of dwelling-based subscriptions required for a telecommunications tower to be profitable if it wereto be constructed in one cell. The key input is the minimum number of dwellings that can financiallysustain the construction of a telecommunications tower through paid subscriptions for wireless broadbandservices.

We assume that 200 dwellings at minimum could support the construction of a broadband tower if afterfive years at least thirty percent of the dwellings with access have subscribed to the technology. Moreover,

after 4 years, there would be a total take rate of 25% and the tower would be paid for after six years. Thereason for such a high cost of the tower is largely due to questions of accessibility within rural and remoteregions. (Sawada et al, 2005)

3.5 AnalysisOnce the irregular shaped EA were converted into regular equal area cells then the raster (grid or cell) calculator canbe used which allows for efficient calculations. At the point the region outlines, dwellings counts, elevation anddistance decay functions are used. We then visit each cell in order to count the number of dwellings that could beprovided with broadband where only dwellings capable of being provided by broadband are counted. There are two

5


6/8

steps to before being able to provide a final raster layer with just dwelling counts. The first step examines a cellsability to establish line-of-site (LOS) communication to each cell. Figure 4 is a 3d model of the LOS process. If acell does have LOS with the currently visited cell (in Figure 4 it is the very centre cell) than it is assumed thatcommunication is possible. Therefore LOS cells are summed together and excluded from being examined in the nextstep, (Figure 6).

The remaining cells within the 50 km radii are examined using a distance decay function developed from the RPA.The distance decay functions used is the one specifically for created for the location of where the cell beingexamined is located. Each cell will have the total number of households that could be provide with broadband.Figure 4b illustrates the final value that would be represented in a cell and this step is repeated two mentioned stepsare repeated for every single cell.

Comm

Line-of-site

Communication

Tower

Figure 4. 3D Visualization of the LOS Examination for a sample map sheet for the Nelson, BC area. Dark patchesrepresent areas visible from a communication tower of 50m.

Figure 5. Diagrams (a - left) I llustration of the cells that could be covered through the examination of LOS from thecentre cell where a dwelling count of the LOS cells are totalled. (b - right) Illustration of the cells that could becovered through the examination of LOS and a distance decay function from the centre cell where a dwelling countof the LOS cells are totalled.

6


7/8

4. Preliminary Resul tsWithout accounting for radio propagation, nearly 85% of Canadian households could be provided with a WiMAXtype wireless connection capable of providing broadband. This study was conducted at varying radii, 5, 10 15, 25,35 and 45 km (Sawada et al. 2005) in order to understand the requirements that will be needed by new wirelesstechnologies. The results are presented inTable 2and Figure 7.

Table 2. Number of dwellings reached with broadband wireless technology at various radii from potentialcommunication towers.

Figure 7. Unserved dwellings that potentially could be provided with fixed wireless broadband internet usingvarious technogies with varying ranges in reach. Does not incorperate a radio propagation analysis.

7


8/8

5. ReferencesDMTI Spatial Inc. Clutter Data. Markham, Ontario. http://www.dmtispatial.com/clutter.htm (Date Accessed: 04

May 2005)Fourie, Denise K. and David R. Dowell. Libraries in the Information Age: An Introduction and Career Exploration.

Greenwood, CO: L ibraries Unlimited, 2002.M. Hata. Empirical Formula for Propagation Loss in Land Mobile Radio Services. IEEE Trans. on Vehicular Tech.,

August 1980, pp. 317-325. [Hata model]Sawada, M., Cossette, D., Chouinard, Grald., Briggs, Gerry., Johnson, P.G. (submitted). A GIS Approach forAssessing Broadband Wireless Market Potential For Rural And Remote Canadian Communities. International

Journal of Knowledge, Science and Society.Thieken, A. Scaling input data by GIS for hydrological modelling. Hydrol. Process. 13, 611-630 (1999).Wellar, B. (1985). "The Significance of Automated Cartography to an Information Society." Cartographica 22(4):

38-50.WiMAX. IEEE 802.16A Standard and WiMAX Igniting Broadband Wireless Access. 2004.

http://www.wimaxforum.org/news/downloads/WiMAX-The_Business_Case-Rev3.pdf (Date Accessed: 20July 2004)

6. AcronymsBroadband High Speed Internet

DA Dissemination AreaDEM Digital Elevation ModelDOI Department of IndustryGIS Geographical Information SystemsIEEE - Institute of Electrical and Electronics EngineersLOS - Line-of-siteMbps Transmission of 1,000,000 bits (megabits) per secondNTDB - National Topographic Data BaseNTS - National Topographic SystemNLOS - Non-Line-of-SiteRF Radio FrequenciesRPA Radio Propagation AnalysisWiMAX Worldwide Interoperability for Microwave Access (IEEE 802.16)MHz - A unit of frequency equal to 1,000,000 hertz

8
http://www.wimaxforum.org/news/downloads/WiMAX-The_Business_Case-Rev3.pdfhttp://www.wimaxforum.org/news/downloads/WiMAX-The_Business_Case-Rev3.pdf

combination of geographic information system (gis) and radio frequency (rf) planning software, to...

Documents